Fear reactions of domestic sheep confronted with either a human or a human-like model.

Fear reactions of ewes towards a human, a human-like model or a control (plastic cylinder 1.95 m high) were studied using a test previously designed and validated, in which 16 behavioural parameters have been interpreted as indicators of fear (e.g. a relatively long time spent away from the stimulus) or absence of fear (e.g. a relatively long time spent eating near the stimulus). In a first experiment, 18 animals were individually confronted with each of the 3 stimuli for 4 min (in a counter-balanced order). The results indicated that both the human and the human-like model elicited greater fear reactions than did the control, and that the former two stimuli did not differ in their fear-eliciting properties. In a second experiment designed to determine if the animals become aware of a difference between the human and the model over time, 32 animals were confronted successively with each of the two stimuli for 8 min. Although there was some habituation towards both stimuli, there was no difference between them even at the end of the test period. These results suggest that a model could be appropriate in experiments involving human presence (at least for a short duration) and would allow for better standardization. It can also be hypothetized that exposing animals to a model would lead to familiarization which might in turn help reduce fear of human beings and thus improve ease of handling and welfare.